Application of Artificial Intelligence in Predicting Earthquakes: State-of-the-Art and Future Challenges

Predicting the time, location and magnitude of an earthquake is a challenging job as an earthquake does not show specific patterns resulting in inaccurate predictions. Techniques based on Artificial Intelligence (AI) are well known for their capability to find hidden patterns in data. In the case of earthquake prediction, these models also produce a promising outcome. This work systematically explores the contributions made to date in earthquake prediction using AI-based techniques. A total of 84 scientific research papers, which reported the use of AI-based techniques in earthquake prediction, have been selected from different academic databases. These studies include a range of AI techniques including rule-based methods, shallow machine learning and deep learning algorithms. Covering all existing AI-based techniques in earthquake prediction, this article provides an account of the available methodologies and a comparative analysis of their performances. The performance comparison has been reported from the perspective of used datasets and evaluation metrics. Furthermore, using comparative analysis of performances the paper aims to facilitate the selection of appropriate techniques for earthquake prediction. Towards the end, it outlines some open challenges and potential research directions in the field.